Agricultural data for life cycle assessments

(1)

Agricultural data for Life Cycle Assessments

B.P. Weidema (ed.) M.J.G. Meeusen (ed.)

Financially supported by: LCA Net Food

Concerted Action PL-97-3079 of the EU Food and Agricultural programme (FAIR)

(2)

(3)

The Agricultural Economics Research Institute (LEI) is active in a wide array of research which can be classified into various domains. This report reflects research within the following domain: ¨ Business development and external factors

(4)

¨ National and international policy issues

(5)

Agricultural data for Life Cycle Assessments Weidema, B.P. and M.J.G. Meeusen

The Hague, Agricultural Economics Research Institute (LEI), 1999

Report 2.00.01; ISBN 90-5242-563-9; Price NLG.120.- (including 6% VAT) 155 p., fig., tab., app.

This book deals with the problem of selection, exchange, and interpretation of agricultural data for use in Life Cycle Assessments. It contains the proceedings of the 2nd European Invitational Expert Seminar on Life Cycle Assessment of Food Products, which was held on the 25 and 26 January 1999 at the Agricultural Economics Research Institute in The Hague. The papers cover the topics: energy consumption, substance balances (especially for nitrogen and phosphorous), and the use of farm typologies and farm accountancy systems for LCA data acquisition.

The discussions and conclusions of the seminar, which are also reported in this book, were moderated by experts on LCA on agricultural products. To complement the topics covered by the seminar, this book contains some invited papers on data for other environmental aspects, such as pesticide use, biodiversity, soil quality, and occupational health. All contributions have been peer re-viewed for acceptance by two or more anonymous reviewers. The first volume of the book consists of chapter A, B, C, (topics), while the second volume of the book deals with chapter D, E (topics).

Orders: Phone: 31.70.3358330 Fax: 31.70.3615624 E-mail: publicatie@lei.dlo.nl Information: Phone: 31.70.3358330 Fax: 31.70.3615624 E-mail: informatie@lei.dlo.nl

Reproduction of contents, either whole or in part: þ permitted with due reference to the source ¨ not permitted

(6)

(7)

(8)

15. The role of the soil in phosphorus cycling

W.J. Chardon 1 Abstract

On a world scale, a net transport of P to the oceans occurs, maintained by weathering of minerals and by erosion. The soil plays a different role regarding phosphorus in different agricultural systems. Without adding fertilisers, a soil can supply P for a limited time, by weathering or by mineralisation of indigenous organic matter. When a higher productivity is aimed at, fertilisers have to be added. In general, more P has to be added than is taken off by the crop to compensate for P becoming less available for plant uptake. In regions with a surplus of animal manure, P contents of the soil can be-come very high, creating problems due to eutrophication of surface waters.

15.1 Introduction

Phosphorus (P) makes up about 0.12% of the earth's crust. It is present in all soils and rocks, in sur-face waters and sediments, and in remains from plants and animals. The world's supply of P comes from mineral deposits, a non-renewable natural resource (Cathcart, 1980). A net transport to the oceans occurs (Tiessen, 1995): the use of mineral P fertilisers in 1990 was estimated as 16 Tg per year. Estimates of yearly P transport to the oceans vary between 21 and 39 Tg year with 23 Tg per year as the best estimate (Howarth et al., 1995). Thus, on a world-wide basis, the P cycle is not closed: the net P transport to the oceans is compensated by weathering of P containing minerals and erosion. However, as will be discussed in section 5, this situation differs strongly between agricultural regions: in many countries, a net input of P to agricultural soils takes place. In surface waters, enrich-ment of P can lead to eutrophication. The symptoms of eutrophication can differ between aquatic systems, and include turbidity, fish mortality, reduction of aquatic macrophytes, and growth of less desirable (toxic) algal species. In most inland waters (rivers and lakes), P is the limiting nutrient for algal growth, which makes reduction of P flow to surface waters important.

15.2 Input of P to soils

Input of P to soils can occur by atmospheric deposition (pollen or dust) and, in agricultural systems, by application of mineral or organic fertilisers.

1

(13)

For Sweden, the amount of P deposition was estimated as 0.07 kg P per ha*year (Sepa, 1993). The P content of dust, mainly originating from wind erosion, will depend on the P content of the soil surface layer in the region where erosion took place.

Several kinds of inorganic fertilisers are used in agriculture. In developing countries, often rock phosphates are used, directly after mining or after acidulation, which releases P otherwise strongly bound to calcium. If applied to soils, rock phosphates will dissolve slowly, and are thus not directly available to crops. In developed countries, the most used fertiliser is superphosphate, which dissolves readily and is thus directly available to crops.

- When organic fertilisers (e.g. compost, animal manure, and sewage sludge) are applied, miner-alisation has to take place before plants can take up P. This will proceed soon after application under wet conditions, with a relatively high ambient temperature, and more slowly at a low temperature or when the organic fertiliser dries out. When the fertiliser is mixed with the soil during or shortly after application, mineralisation will proceed more quickly.

15.3 Soil processes

After application of inorganic P fertilisers to a soil, or when P has been released from organic fertilis-ers by mineralisation, several reactions with the soil can take place:

- adsorption: fast reaction of P with the outside of soil particles; - absorption: slow migration of P into the pores of soil aggregates;

- immobilisation: P is incorporated into soil organic matter; this is especially important on grass-land soils, where a strong accumulation of organic P can be found;

- precipitation: binding of P with other chemical elements, e.g. calcium; this only occurs in soils with a high pH.

Adsorption of P mainly occurs onto hydroxides of iron or aluminium. The amount of these compounds in soils is limited. When the capacity is nearly used up, both the availability of P for up-take by plants and the possibility of P loss to the environment increases. In general, more P has to be added than is taken off by the crop to compensate for P becoming less available for plant uptake. Absorption (migration of P into soil aggregates) is considered to be the main process responsible for what can be called 'inevitable loss of P'.

15.4 Loss of P from soil

As will further be discussed in the paper of Heathwaite (this volume), loss of P from soils can pro-ceed via different pathways:

(14)

- transport over the soil surface with water from rainfall or snow melt, when the infiltration ca-pacity of the soil is exceeded ('surface runoff') and the soil has a certain slope. This is a fast process, and the P content of the receiving surface water is raised shortly after rainfall occurs; - transport via highly permeable parts of the soil ('preferential water flow'). This is also a fast process, leading to a quick rise of P in surface water at the start of rainfall. It may occur when: - clay soil contains cracks;

- organic soils are dry; - the soil is artificially drained;

- parts of the soil are water repellent; or - the soil contains permanent wormholes.

The transport routes mentioned above differ from an environmental point of view. When water transport occurs via the soil matrix, soil P content near the water table determines the environmental risk. In the case of surface runoff or preferential flow, the P content of the soil's upper layer is more important, implying that after excessive P application to the soil surface environmental problems by P loss can be expected in an early stage.

15.5 Balance of P

The difference between the amount of P applied to a field or on a farm via fertilisers, and the amount of P exported via crops, milk or meat, can be called the P balance (see also the paper of Withers in this volume). In developing countries, when farmers cannot afford the use of fertilisers, the balance can be negative: P in the soil becomes depleted, and yields are generally low. For sub-Saharan Af-rica, an average P loss of 2.5 kg P per ha*year was calculated (Stoorvogel et al., 1993). The P taken up by crops originates from weathering or mineralisation of indigenous organic matter. Often, after several years of exploitation, the soil will be left and new land will be cultivated.

When the yield pursued is higher, fertilisers have to be applied in order to build a reserve of P in the soil, and the P balance will be positive. This is the case in most developed countries. The situation changes dramatically when a farm (or region) has a limited soil surface area, a relatively large production of P containing manure exceeding crop offtake, and no possibilities to export the manure from the farm or region. The balance can become strongly positive through the purchase of P via ani-mal feed. In this case, the aim of P application to the soil is not maintenance of soil fertility, but the disposal of the manure produced. This will lead to a fast build up of a P reserve in the soil, far ex-ceeding crop needs, and creating a risk for the environment. For different European countries, the national P surplus was calculated; results for 1992 are given in table 15.1. It has to be kept in mind that also in countries with a low or moderate P surplus, regions may exist where confined animal pro-duction is concentrated, and a much larger surplus can be found. Regional differences are illustrated for the United Kingdom, France, and Spain.

(15)

It will be clear that P surpluses as shown for e.g. Belgium and the Netherlands will create, or already have created, environmental problems due to P loss. A positive aspect is that the use of chemical P fertilisers and P surpluses tend to decrease, as is shown in table 15.2.

Although the calculated surpluses of P have decreased in most countries, with a positive result of the P balance, the P reserves in the soils still increase.

Table 15.1 Surplus of total-P (kg P per ha*year) for the average farm of some European countries and re-gions in 1990/1991

Country Region Country Region

United Kingdom 6 Ireland 15

England West 8 Switzerland 16

Scotland 3 Germany 21 Denmark 8 France 28 Spain 12 Bretagne 37 Galicia 20 Limousin 13 Extremadura 10 Belgium 36 Greece 15 Netherlands 40

Source: Data from Brouwer et al. (1995).

Table 15.2 Reduction (%) in application rates of chemical P fertilisers and P surplus for some European countries

Country Period, 1985- Fertiliser Period, 1985- Surplus

Belgium 1992 24 1992 5 Denmark 1992 29 1990 0 France 1992 19 1990 0 Germany 1993 58 1992 67 Netherlands 1992 11 1992 19 Norway 1992 18 1992 42 Sweden 1992 50 1990 20 Switzerland 1990 7 1990 20 United Kingdom 1993 17 1993 17

Source: Data from De Walle and Sevenster (1998).

(16)

to an increased risk of P transport to surface water, on a short term in case of surface runoff or pref-erential flow, and on a longer term in case of matrix flow.

(17)

References

Brouwer, F.M., F.E. Godeschalk, P.J.G.J. Hellegers and H.J. Kelholt, Mineral balances at farm level in the

Euro-pean Union. Research Report 137, LEI-DLO, The Hague, the Netherlands, 141 p., 1995.

Cathcart, J.B., 'World phosphate reserves and resources', pp. 1-18, In: F.E. Khasawneh et al.(ed.), The role of

phos-phorus in agriculture. ASA, CSSA, SSSA, Madison USA, 1980.

Howarth, R.W., H.S. Jensen, R. Marino and H. Postma, 'Transport to and processing of P in near-shore and oce-anic waters', pp. 323-345, In: H. Tiessen (ed.) Phosphorus in the global environment. Transfers, cycles and mana-gement. Wiley, 1995.

SEPA, Eutrophication of soil, freshwater and the sea. Report 4244, Swedish Environ. Protection Agency, Solna, Sweden, 207 p., 1993.

Stoorvogel, J J, Smaling E M A, Janssen B H. Calculating soil nutrient balances in Africa at different scales: 1. Supra-national scale. Fert. Res. 35:227-235, 1993.

Tiessen, H., 'Introduction and synthesis', pp. 1-6, In: H. Tiessen (ed.) Phosphorus in the global environment. Transfers, cycles and management. Wiley, 1995.

Walle, F.B. de, J. Sevenster, Agriculture and the environment. Minerals, manure and measures. Kluwer, Dordrecht. 211 p., 1998.

Further reading

Information on world P reserves, fertilizer manufacturing and use, reactions of P in soil, and plant nutrition can be found in:

Khasawneh, F.E., E.C. Sample and E.J. Kamprath, The role of phosphorus in agriculture. ASA, CSSA, SSSA, Madison USA, 910 p., 1980.

General reviews on P cycling can be found in:

H. Tiessen (ed.), Phosphorus in the global environment. Transfers, cycles and management. Wiley, 462 p., 1995. Recent reviews on the relation between agriculture, soil P and eutrophication are given in:

Tunney, H., O.T. Carton, P.C. Brookes and A.E. Johnston (eds), Phosphorus loss from soil to water. CAB Interna-tional 1997, 467 p., 1997.

(18)

16. Cycling and sources of phosphorus in agricultural

systems and to the wider environment: a UK perspective

P. J. A. Withers 1 Abstract

Inputs of phosphorus (P) in fertilisers and feeds often exceed P exports in harvested produce on in-tensively managed agricultural holdings, especially those operating with high livestock densities. The build-up of surplus P in the soil, together with frequent spreading of relatively large amounts of recy-cled excretal P on farms, are of environmental concern with respect to the transfer of P in land run-off to water causing eutrophication.

The frequency of P application, the amount of the P surplus and the soil depth over which sur-plus P is distributed varies considerably between different regions and farming systems with implications for P transfer. Trends towards continuous cultivation, slurry based livestock systems and the installation of tile under-drainage in arable and grassland systems are also considered to have in-creased the ease with which soil-accumulated and freshly-applied P are lost to surface waters. Accelerated P losses are not derived equally over the catchment area, and may originate only from fields with inherently high P loss risk, or which are mismanaged.

Climate, landscape, soil type, farming system and farm management data are required to define the transfer of soluble P and P associated with eroding soil particles to watercourses, but these show large regional variation due to the wide distribution of soil parent materials, climate and topography affecting natural P loss, and diverse patterns of farming systems with regard to land use, P inputs and land management. Expert systems are required to compare the relative importance of regional differ-ences in site and agricultural management factors in order to quantify the P emissions associated with regionally produced agricultural products. In some areas, some form of control over agricultural P inputs, and/or the transport of P within the landscape, is, or will be, required in future to help maintain water quality for a range of uses.

Agricultural crops require adequate amounts of phosphorus (P) for healthy growth and to maximise the utilisation of other nutrients, especially nitrogen (N). Similarly, livestock require adequate amounts of P in their diet to prevent against deficiencies, which might impair their health and performance. As a consequence, P fertilisers and minerals have been routinely and liberally imported onto farms in re-sponse to economic and political pressures to maximise agricultural production. Within the developed countries, farming systems have generally become more intensive, with a greater proportion of land

1

ADAS Bridgets Research Centre, Martyr Worthy, Winchester, Hampshire SO21 1 AP, UK. Tel.: +441962792424; Fax.: +441962779739; e-mail: Paul_Withers@ADAS.co.uk

(19)

under cultivation, with underdrainage and/or with increased animal densities. For example, agricultural census returns indicate that the numbers of animals kept on specialist livestock holdings in the UK typically increased 3-fold between 1965 and 1993.

Increased P imports on intensively farmed holdings, particularly livestock holdings, have led to a greater reliance on readily-available inorganic fertiliser and feed P products, larger amounts of faecal P requiring disposal onto land and an accumulation of surplus P in the soil, of which a greater proportion is in readily-exchangeable form (Isermann, 1990; Brouwer et al., 1995; Tunney et al., 1998). These trends have raised a number of environmental concerns; the mining of exhaustible rock phosphate reserves in developing countries and the air pollution associated with manufacture of inor-ganic P products; the accumulation of potentially harmful metals in soils from the repeated application of rock phosphate and its products, particularly cadmium (Cd), and the accelerated loss of freshly-applied and soil-accumulated P from agricultural land to water causing eutrophication (Withers and Sharpley, 1995). The relative importance of these concerns varies between different countries, but perhaps, the most widespread and increasing problem is that associated with eutrophication.

The extent of eutrophication problems in freshwaters is most commonly related to P inputs and only very low concentrations of P are required for eutrophication symptoms to appear (Gibson, 1997). The role of agriculture in the eutrophication process has rarely been clearly defined, largely because anthropogenic sources are usually the major source of P loads, and P losses in land run off are difficult to quantify due to their diffuse nature. They emanate from a number of source areas within the landscape, and their amount, form and timing are very variable as a result of short-term and often unpredictable changes in hydrological conditions and farming practices; rotational cropping, the appli-cation of fertilisers and manures, or the movement of animals from one field to another (Lennox et al., 1997). Recent monitoring of rural catchments suggests that the loads and concentrations of P in land run-off are sufficient to cause eutrophication and that they have increased under intensification (Foy and Withers, 1995; Heathwaite et al., 1996). This paper reviews the cycling of P within agri-cultural systems and the associated risks to environmental life cycles.

16.2 Phosphorus cycling within agriculture

16.2.1 Fertilisers and feeds

Unlike N, P is a conservative element whose inorganic forms become strongly bound to soil colloids. The degree of binding depends on the nature of the adsorbing surface and the ionic composition of the soil solution, but essentially P is relatively immobile in soil. Field experiments have consistently demonstrated that the proportions of fertiliser and manure P utilised by crops in the year of applica-tion is low (< 20%). Since crop P requirement is largely derived from the soil, it is the ease of exchangeability of soil P, as assessed by standard soil extraction tests, which forms the basis of

(20)

fertil-reserves in the soil reach a certain level; C1 in figure 16.1. Above this critical level, P inputs need only to match crop P offtake, except perhaps for some soils where P becomes progressively unavailable to crops due to fixation processes (Withers et al., 1994; Bertilsson and Forsberg, 1997). On soils of adequate P fertility, the availability of P in different fertilisers and manures is therefore not significant in terms of crop production, and it is the total P input that must be regulated to prevent the excessive soil P accumulation which leads to accelerated P transfer to water (figure 16.1).

> output ≡≡ output < output omit

Lost yield

build-up maintain reduce control

Accelerated loss C₁ C₂ Total P inputs (kg ha-1₎ Harvest yield (t ha -1 ) P loss (kg ha -1) Readily-exchangeable soil P (mg kg-1₎

Figure 16.1 Conceptual diagram of critical concentrations of readily-exchangeable P for optimum crop yield (C1) and accelerated P loss to water (C2) in relation to P input strategies

Within animal production, the apparent efficiency of utilisation of feed P inputs is also low, with typically 70-80% of fresh P intake excreted in dung and urine (Lynch and Caffrey, 1997). Much of the dietary P intake of livestock on intensive holdings originates from imported concentrates, which are generally formulated to insure against deficiencies of P in other constituents of the diet and to overcome genetic variability in P absorption between animals. Feed compounders have, therefore, been less concerned with matching the actual P requirements of animal production, but have tended to include inorganic P supplements in generous quantities to avoid the possibility of reduced animal health and fertility. In a recent systems study in the UK, a 40% reduction in imported P fed to dairy cows yielding >6,000 l per year did not appear to impair milk production compared to a conventional dairy herd receiving commercially formulated inputs (table 16.1). Recent reviews indicate there may be scope to reduce P inputs, and/or improve P utilisation, through better dietary manipulation, since

(21)

it is now well established that P fed in excess of dietary requirements is simply excreted without im-proving production (Valk et al., 1998; Damgaard-Poulson, 1998). Consequently, farmers have little or no idea of the amount of P ingested by their animals, or the extent to which it is being over-supplied. This is in direct contrast to the planning of fertiliser inputs for crop production. Unlike crop-ping systems on fertile soils, the P availability of individual feed ingredients is important to utilisation, especially for non-ruminants, which cannot utilise phytate-P effectively.

16.2.2 Surplus phosphorus

The relative proportions of feed and fertiliser imports compared to P output in farm produce deter-mines the amount of surplus at the scale measured, and the degree of soil P accumulation. Within the UK, P imports exceed P exports by about 10 kg P per hectare, averaged over the total agricultural land area (table 16.1).

Table 16.1 Inputs and outputs of phosphorus (kg per ha) in the UK and in different UK farming systems

UK Dairy systems b) Arable c) Upland d) (1993)  (1985-98) (hill sheep) a) high output reduced loss

Livestock density (LU ha-1) - 1.9 1.6 -Inputs Atmosphere 0.3 0.5 0.5 0.5 0.1 Imported feeds/minerals 2.6 23.3 13.4 - 0.2 Imported fertilisers 9.3 13.6 0.8 24.8 0.5 Sewage sludge 0.4 - - - -Sub-total 12.6 37.4 14.7 25.3 0.8 Outputs

Milk and eggs 0.7 11.6 10.2 -

-Meat and wool 1.7 0.8 0.7 - 0.1

Grain and straw 0.6 - - 16.7 0.1

Misc e) - 1.7 1.1 -

-Sub-total 3.0 14.1 12.0 16.7 0.2

Surplus 9.6 23.3 2.7 8.6 0.6

Losses 0.26 f) 0.36 f) 0.3

a) Withers (1996); b) Results from a study comparing an intensive high output dairy farming system with a dairy system receiving reduced P inputs in feeds and fertilisers and reduced stocking density after Withers et al. (1999); c) Commercial farm taken over in 1985 with detailed records of fertiliser P inputs and crop yields (Whinfield pers.

(22)

The largest P import is fertiliser, without which the UK P surplus would be close to zero. How-ever, P fertilisers are required in areas, which have no livestock, and it is probably the uneven redistribution of faecal P that is responsible for the P imbalance. Indeed, the exports of crop and ani-mal products represent only about one quarter of the c. 210 106 kg of P recycled in UK agriculture either as excreta, or as home-grown feedstuffs fed to livestock (Withers, 1996). The redistribution of these large P amounts, which often require handling on more than one occasion, increases the op-portunity for wastage and P loss at the farm scale. Comparing overall P inputs to outputs, UK agriculture would appear to be only 24% efficient (table 16.1). Most of the developed countries op-erate a national P surplus (Brouwer et al., 1995).

At the farm and field scale, surpluses of P may differ substantially from the national picture de-pending on the type of farming system (table 16.1). For example, upland grass/hill farms may operate a very small surplus compared to an intensive dairy farm with significant feed imports, although their P loss rates may be very similar. Unlike N, the link between the P surplus and the loss of P is not clear due to differences in the patterns of flow and retention of P in the soil. Arable farms without access to manure inputs may either be in balance, or be in surplus. For example the arable farm in table 16.1 has been trying to build-up soil P fertility with fertiliser P inputs in excess of P offtake, and has a surplus similar to the UK average. Horticultural holdings use relatively large amounts of fertiliser and manure P in their multi-cropping systems and often have high concentrations of readily-exchangeable P in the soil.

The greatest risk of P surplus is on farms, which generate or import far more manure than can be sensibly recycled to the available land area. For example, Brouwer et al. (1995) calculated an average P surplus of 269 kg per ha*year for granivore farms where stocking densities are very high and there is little land area for recycling of the manure. In addition, national surveys show that those farmers who do recycle manures on their farms do not take account of their fertiliser value, but still apply substantial amounts of inorganic P fertiliser (Edwards et al., 1997), a practice which limited research suggests is unnecessary (Smith and Van Dijk, 1987; Van Dijk and Sturm, 1983). Hence, the frequency of P application, or deposition for grazing animals, and the potential for large P sur-pluses, on individual fields is locally very variable. Fundamental differences in the distribution of surplus P with soil depth also exist between uncultivated and regularly cultivated land. Since the amount of P loss in storm run-off has been shown to be related to the rain-soil interaction within a shallow (1-2 cm) surface layer (Ahuja, 1986), differences in the distribution of soil P between arable land and grassland maybe environmentally significant.

16.3 Diffuse phosphorus loss from agricultural land to water

16.3.1 Sources and transfer of phosphorus

Since nutrient transport from land to water is a natural process, it implies that agriculturally driven P limiting eutrophication problems probably occur due to a sustained acceleration of diffuse P loss in land run-off (Gibson, 1997). Agriculturally derived point source P inputs, such as stormflow from

(23)

farm waste stores or farmyards, may also contribute significant P loading to watercourses, but proba-bly do not re-occur on a regular or widespread basis. Concentrations of P in rainfall can be large; for example Withers et al. (1999) recorded values > 700 µg per l associated with wind-blown soil particles in a dry arable area, but the loads landing directly on watercourses are comparatively small. Hence, agricultural P loss largely occurs diffusely in land run-off due to the transport of soil particles (erosion) and in soluble form (surface run-off and leaching; figure 16.2). Freshly-applied fertilisers and manures, including those deposited by grazing animals, which are not incorporated into the soil also contribute directly to incidental PP and DP loads in run-off producing land areas, especially on grassland (Heathwaite, 1997).

Figure 16.2 Sources, processes, forms and pathways of P loss from agricultural land to water

Analytically, the transfer of P in land run-off has been separated into that which occurs in asso-ciation with soil particles passing >0.45µm (particulate P or PP) and that which occurs in soluble form (DP, <0.45µm). In reality, amounts of P in true solution may be very small, and most P is probably in particulate or colloidal form (Edwards and Withers, 1998). Although often viewed as a

Fertilisers/manures Soil phosphorus Particulate (>0.45µm) ⇔ Soluble (<0.45µm) Groundwater Drainflow Surface run-off Detachment Dissolution Detachment Sedimentation Dissolution Leaching

(24)

of connectivity to the stream include ditches, tracks and roads (Harden, 1992). The relative propor-tions of PP and DP in land run-off therefore depend on the complex interaction between climate, topography, soil type, soil P content, type of farming system, and farm management. The influence of agriculture and agricultural practices on the potential for P loss under the site hydrological condi-tions, and the identification of effective P loss control strategies, requires knowledge both of P inputs (nutrient management) and P transport (land management) (Withers and Jarvis, 1998).

16.3.2 Nutrient management

Phosphorus accumulation in the soil increases all P fractions, but is accompanied by an increase in the ratio of exchangeable to total P. Linear and non-linear relationships between readily-exchangeable P (soil test P) and DP loss in surface and sub-surface run-off have been recently dem-onstrated in laboratory studies (Sharpley, 1995), plot studies (Heckrath, et al., 1995), and/or whole field comparisons (Smith et al., 1998). Such data suggests there is a critical soil test P level above which DP loss is greatly accelerated, or becomes unacceptable; C2 in figure 16.1. Differences in the

nature and/or slope of the relationship depend on soil type, land management and the depth and time of interaction between stormflow and the soil. Although current soil agronomic tests may not be the most appropriate method for assessing P loss potential in storm run-off, it is usually the only data which is normally routinely available on a regular basis to quantify soil P accumulation and P loss risk. However, the analytical methods used differ widely between different countries and comparison of soil test P concentrations within the EU is problematical (Tunney et al., 1997).

Impacts of total P accumulation on PP loss in storm run-off are more difficult to quantify and have not been extensively studied. Erosion is a selective process and particles which are transported long distances in run-off tend to be very fine-textured, highly P reactive and enriched with P com-pared to the bulk soil (Sharpley and Smith, 1990; McGuire et al., 1998). Annual P surpluses may only represent a very small proportion of the total soil P content, but depending on the soil depth over which P accumulation is measured. For example, a UK surplus of 10 kg P ha-1 is only 0.4% of the median content of total P down to a plough depth of 25 cm, and may need to be continued for many years before increases in PP transport could be detected on arable soils. The relative contribution of agriculturally-derived P in the soil compared to the native P content under non-agricultural use is therefore unclear. Some parent materials (for example, chalks and limestones) are naturally rich in total P and these must be considered as natural hotspots of P. On other soils, surplus P inputs may greatly influence soil total P, especially since P inputs are taken up preferentially by fine aggregates (McGuire et al., 1998). There is currently no soil test available that quantifies the potential for par-ticulate P loss from a given soil type.

When fertilisers and manures are applied to the surface of the soil, there is also a risk of inci-dental P loss in storm run-off, especially if applied to soils already at field capacity, to frozen soils, or to cracked or recently underdrained soils. Although the amounts of P lost are generally very small (<5%) in relation to the total P amounts applied, the concentrations are well above those required for eutrophication to occur, with up to 30 mg per l recorded in field experiments. Measured losses depend on the rate, time, method and frequency of application, the form in which the P is applied,

(25)

hydrological conditions following application, and amounts of vegetative cover (Heathwaite, 1997; Smith et al., 1998). These factors are fundamentally different between arable and grassland farming systems. Regional information on the rate and type of fertiliser/manure and the method and timing of application is needed in relation to site factors, but it is unclear how widely available these are. In the UK, this information is derived by an annual survey of fertiliser practice (Burnhill et al., 1997). 16.3.3 Land management

Erosion is the main process of P transfer from agricultural land to water and is considered to have increased as a result of modern farming techniques. Field experiments and erosion surveys indicate that P export rates from arable land typically range from 0.1-30 kg P per hectare (Schonning et al., 1995; Chambers, 1997). In the UK, major contributing factors are the increase in the area sown to winter cereals, the introduction of tramlines which concentrate and increase the velocity of water flow, the removal of hedgerows which increases the length of slope, and the reduced soil stability arising from continuous cultivation (Evans, 1990; Spiers, and Frost 1985). For example, in a survey of 145 eroding fields in England and Wales monitored during 1989-1994, 80% were cropped to winter ce-reals (Chambers, 1997).

However, the development of rills and gullies at the surface of cultivated fields is not the only form of erosion from agricultural fields. Significant losses of particulate P also occur without any obvi-ous disturbance of the soil surface, in sub-surface flow through tile drains and from poached grassland (Haygarth et al., 1998b; Heathwaite, 1997; Hodgkinson and Withers, 1996). For example, Hodg-kinson and Withers (1996) measured particulate P losses of ca. 1.5 kg per hectare from a field drain in a dispersive silty soil under arable cropping. Using 137Caesium fingerprinting techniques, Grant et al. (1996) indicated that transported sediment-bound P in field drains originated from the topsoil and travelled down soil macropores and fissures. Heathwaite et al. (1990) found significantly greater loss of sediment bound P under high stocking densities compared to low stocking densities on grassland in a high rainfall area. Gateways and drinking troughs are high risk source areas for particulate P loss due to the effects of heavy poaching reducing ground cover and creating soil disturbance.

Quantification of particulate P emissions requires data on the amount of suspended sediment in run-off and the P concentration of the suspended sediment. Generally, transported particulate P is composed of silt and clay fractions, which are enriched in P compared to the bulk soil (Sharpley, 1985; Sharpley and Smith, 1990). The enrichment ratio will vary between sites (<1-5) according to soil type, P fertilisation history and the depth of P accumulation, but has been shown to be related to soil loss within a site (Sharpley, 1985). Models are available to predict these two parameters but they require validation at a wide range of field and catchment scale. In particular, there is little data on the differences in the spatial distribution of sediment-bound P within catchments of different size.

(26)

16.4 Conclusions

Recent research indicates that accelerated losses of P from intensively managed farmland to water may contribute to eutrophication problems within catchments. However the major source areas and pathways of diffuse P loss, and the specific contribution of agriculture or agriculture practices are dif-ficult to identify and quantify accurately. There is marked spatial and temporal variation in the loads and concentrations of P in land run-off due to the complex interaction between the amount and inten-sity of rainfall, the susceptibility of the soil to detachment and erosion, the soil P content and its degree of saturation and the presence of fertilisers, manures, and crop residues at the soil surface.

Undesirable agricultural practices which accelerate PP and/or DP loss are the unnecessary accumulation of P, the application of P amendments at rates and at times which cause direct run-off, and the adoption of land management practices, which increase erosion risk on unstable soils. Other farm management practices, which have been introduced to improve agricultural production (in-creased winter cereals, tramlines, underdrainage, slurry-based livestock systems), probably also increase the potential for diffuse P loss but have other advantages, which makes their desirability more difficult to evaluate. Fundamental differences in the amount and frequency of P application, the amount of surplus P, the depth of soil P accumulation, inherent ground cover and hydrological condi-tions exist between cultivated and uncultivated systems which have implicacondi-tions for P transfer to water.

Although the precise impacts of diffuse P loss from agricultural land on water quality are poorly understood, it is clear that certain farm management practices can cause greatly accelerated P loss in land roff, which has the potential to cause eutrophication problems and therefore are both un-necessary and unacceptable. Methodologies need to be established to quantify the total P load derived from agriculture, the impact of this load (or concentration) on the biotic equilibrium at critical times of the year, and to pinpoint the major source areas and pathways contributing the P loss so that control options can be implemented effectively.

Quantification of the P loss risk associated with regionally produced agricultural products re-quires assessment of the relative importance of regional differences in site and agricultural management factors. The availability of regional data is likely to vary considerably between different countries requiring inventories of such data within broad ecological zones. This information is also needed to highlight where in the production cycle most P emission occurs and how it might be con-trolled, since causal factors may be very different in different regions both within and between countries.

References

Ahuja, L.R., Characterization and modelling of chemical transfer to runoff. Advances in Soil Science 4:149-188, 1986.

Bertilsson, G.C. and C. Forsberg, 'Sustainable phosphorus management in agriculture', pp. 273-282, In: Tunney H, Carton O T, Brookes P C, Johnston A E (Eds), Phosphorus Loss from Soil to Water, Wallingford, CAB Interna-tional, 1997.

(27)

Brouwer, F.M., F.E. Godeschalk, P.J.G.J. Hellegers and H.J. Kelholt, Mineral Balances at the Farm Level in the

European Union. LEI-DLO, The Hague, 1995.

Burnhill, P., A.G. Chalmers and L. Owen, Fertilizer Use on Farm Crops for Crop Year 1996. British Survey of Fer-tilizer Practice. Edingburgh, The Stationary Office, 1997.

Chambers, B.J., Guidelines to reduce the risk of water erosion on susceptible soils used for arable cropping. MAFF Final Report for Project SP0401. Ministry of Agriculture, Fisheries and Food, London, 1997.

Damgaard-Poulsen, H., 'Phosphorus utilization in pigs: a review', 12 p. (Abstract), In: Practical and Innovative

Measures for the Control of Phosphorus Loss to Water, Proceedings of an International Workshop, 16-19 June,

Belfast, Northern Ireland, 1998.

Dijk, T.A. van and H. Sturm, Fertilizer value of animal manures on the continent. Proceedings No. 220. York, The Fertilizer Society, 1983.

Edwards, A.C. and P.J.A. Withers, A Soil phosphorus management and water quality: a UK perspective. Soil Use and Management 14:124-130 (Supplement), 1998.

Edwards, A.C., P.J.A. Withers and T.J. Sims, Are current fertilizer recommendation systems for phosphorus

ade-quate? Proceedings of the Fertilizer Society No. 404. York, The Fertilizer Society, 1997.

Evans, R., Water erosion in British farmer's fields - some causes, impacts, predictions. Progress in Physical Geog-raphy 14:199-219, 1990.

Foy, R.H. and P.J.A. Withers, The contribution to agricultural phosphorus to eutrophication. Proceedings of the Fertilizer Society No. 365. York, The Fertilizer Society, 1995.

Gibson, C.E., 'The dynamics of phosphorus on freshwater and marine environments', pp. 119-135, In: H. Tunney,

Carton O T, Brookes P C, Johnston A E (Eds), Phosphorus Loss from Soil to Water, Wallingford, CAB

Interna-tional, 1997.

Grant, R., A. Laubel, B. Kronvang, H.E. Andersen, L.M. Svendsen and A. Fuglsang, Loss of soluble and

particu-late phosphorus from arable catchments by subsurface drainage. Water Research 30:2633-2642, 1996.

Harden, C.P., Incorporating roads and pathways in watershed-scale hydrologic and soil erosion models. Physi-cal Geography 13:368-385, 1992.

Haygarth, P.M., P.J. Chapman, S.C. Jarvis and R.V. Smith, Phosphorus budgets for two contrasting grassland

farming systems in the UK. Soil Use and Management 14:160-167 (Supplement), 1998a.

Haygarth, P.M., L. Hepworth and S.C. Jarvis, Forms of phosphorus transfer in hydrological pathways from soil

under grazed pasture. European Journal of Soil Science 49:65-72, 1998b.

Heathwaite, A.L., 'Sources and pathways of phosphorus loss from agriculture', pp. 205-223, In: Tunney H, Carton

(28)

Hodgkinson, R.A. and P.J.A. Withers, 'Non-point source phosphorus loss from small agricultural catchments', pp. 120-126, In: Petchey A M, D'Arcy B J, Frost C A (Eds), Diffuse Pollution and Agriculture, Aberdeen, Scottish Agricultural College, 1996.

Isermann, K., Share of agriculture in nitrogen and phosphorus emissions into the surface waters of Western

Europe against the background of their eutrophication. Fertiliser Research 26:253-269, 1990.

Lennox, S.D., R.H. Foy, R.V. Smith and C. Jordan, 'Estimating the contribution from agriculture to the phosphorus load in surface water', pp. 55-75, In: Tunney H, Carton O T, Brookes P C, Johnston A E (Eds), Phosphorus Loss

from Soil to Water, Wallingford: CAB International, 1997.

Lynch, P.B. and P.J. Caffrey, 'Phosphorus requirements for animal production', pp. 283-296, In: Tunney H, Carton O.T., Brookes P.C., Johnston A.E. (Eds), Phosphorus Loss from Soil to Water, Wallingford, CAB International, 1997.

McGuire, R.O., A.C. Edwards and M.J. Wilson, Influence of cultivation on the distribution of phosphorus in three

soils from NE Scotland and their aggregate size fractions. Soil Use and Management 14:147-153 (supplement),

1998.

Schonning, P., E. Sibbessen, A.C. Hansen, B. Hasholt, T. Heidmann, M.B. Madsen and J.D. Nielsen, Surface

run-off, erosion and loss of phosphorus at two agricultural soils in Denmark - plot studies 1989-92. SP Report No.

14, Danish Institute of Plant and Soil Science, Foulum, Denmark, 196 p., 1995.

Sharpley, A.N. 'The selection erosion of plant nutrients in runoff'. In: Soil Science Society of America Journal

49:1527-1534, 1985.

Sharpley, A.N. 'Dependence of runoff phosphorus on extractable soil phosphorus'. In: Journal of Environmental

Quality 24:920-926, 1995.

Sharpley, A.N. and S.J. Smith, 'Phosphorus transport in agricultural run-off: the role of soil erosion', pp. 351-366, In: Boardman J, Foster L D L, Dearing J A (Eds), Soil Erosion on Agricultural Land, Chichester, John Wiley and Sons, 1990.

Smith, K.A. and T.A. van Dijk, 'Utilisation of phosphorus and potassium from animal manures on grassland and forage crops', pp. 88-102, In: H.G. Van De Meer et al. (Eds), Animal Manure on Grassland and Fodder Crops, Dordrecht, Matinus Nijhoff Publishers, 1987.

Smith, K.A., A.G. Chalmers, B.J. Chambers and P. Christie, Organic manure phosphorus accumulation, mobility

and management. Soil Use and Management 14:154-159, 1998.

Spiers, R.B. and C.A. Frost, The increasing incidence of accelerated soil water erosion on arable land in the

East of Scotland. Research and Development in Agriculture 2:161-167, 1985.

Stamm, C., H. Fluhler, R. Gachter, J. Leuenberger and H. Wunderli, 'Preferential transport of phosphorus in drained grassland soils'. In: Journal of Environmental Quality 27:515-522, 1998.

Tunney, H., H. Breeusma, P.J.A. Withers and P.A.I. Ehlert, 'Phosphorus fertilizer strategies: present and future', pp. 177-203, In: Tunney H, Carton O T, Brookes P C, Johnston A E (Eds), Phosphorus Loss from Soil to Water, Wallingford: CAB International, 1997.

Tunney, H., R.H. Foy and O.T. Carton, 'Phosphorus inputs to waters from diffuse sources', pp. 25-39, In: Wilson

(29)

Valk, H., J.A. Metcalf and P.J.A. Withers, 'Prospects for minimising phosphorus excretion in ruminants by dietary manipulation', pp. 10 (Abstract), In: Practical and Innovative Measures for the Control of Phosphorus Loss to

Water, Proceedings of an International Workshop, 16-19 June, Belfast, Northern Ireland, 1998.

Withers, P.J.A. Phosphorus cycling in UK agriculture and implications for water quality. Soil Use and Manage-ment 12:221 (Abstract), 1996.

Withers, P.J.A. and S.C. Jarvis, Mitigation options for diffuse phosphorus loss to water. Soil Use and Manage-ment 14:186-192, 1998.

Withers, P.J.A., S. Peel, R.M. Mansbridge, A.C. Chalmers and S. Lane, Transfers of phosphorus within three dairy

farming systems receiving varying inputs in feeds and fertilizers. Nutrient Cycling in Agroecosystems (In press),

1999.

Withers, P.J.A. and A.N. Sharpley, 'Phosphorus Fertilizers', pp. 65-107, In: J E Rechcigl (ed), Soil Amendments and

Environmental Quality,. Boca Raton, Lewis Publishers, 1995.

Withers ,P.J.A., R.J. Unwin, J.P. Grylls and R. Kane, 'Effects of withholding phosphate and potash fertilizer on grain yield of cereals and on plant-available phosphorus and potassium in calcareous soils'. In: European Journal

(30)

17. Flows of phosphorous in the environment: identifying

pathways of loss from agricultural land

Louise Heathwaite 1 Abstract

Phosphorus (P) flows in the environment are hydrologically driven, and enable the transport of poten-tially mobile P from agricultural land to receiving waters. Two key modes of transport exist: surface runoff and subsurface flows. While surface runoff remains an important pathway of P loss, recent research demonstrates the potential for subsurface transport of P in macropore flow and from drained land. The forms of mobilised P differ according to the transport pathway. For grassland, dissolved P is transported in surface runoff but particulate P is proportionately more important in macropore and drainflow - especially during storm events. Tilled land generally shows high particulate P trans-port. Where livestock intensification has increased the rate of manure returns to land, there is clear evidence of enhanced P transport, both as incidental losses in surface runoff and through matrix or preferential flow in subsurface pathways.

Abbreviations: TP total P; TPP total particulate P; POP particulate organic P; PIP particulate inorganic P (adsorbed onto Fe/Al complexes and as Ca/Mg phosphate); TDP total dissolved P; DIP dissolved inorganic P (molybdate reactive P); DOP dissolved organic P (may include P oxides).

The flow of phosphorus (P) from agricultural land depends on the coincidence of source and trans-port controls. Phosphorus source areas have a high potential to contribute P; they are often spatially limited and may include land of high soil P status or reflect agricultural land uses which increase sur-face P concentrations, for example, intensively grazed grassland or certain arable crops. Phosphorus source areas are dynamic and reflect agricultural land use and management. Transport factors de-scribe the hydrological processes, which translate P source areas into P loss areas. Not all catchment areas are equally vulnerable to P loss; certain areas contribute runoff (both surface and subsurface) more readily than others do. For example, hillslope hollows become saturated through the confluence of subsurface water with the consequent rise in the local water table and increased risk of saturation-excess surface runoff (see later). In terms of P transport, such areas do not pose a risk unless they are coincident with P source areas. This means that within an agricultural catchment it is possible to have areas with a high potential to contribute P but no P transport if the hydrological connectivity does not exist; conversely we may have areas with high hydrological connectivity but no P transport because they do not link to P source areas. This paper will examine the hydrological pathways of P

1

(31)

transport in an attempt to account for their significance in contributing to P flows from agricultural land to receiving waters.

Various conceptual models have been developed to provide an overall representation of the mechanisms of P export from agricultural land. Heathwaite (1995) suggests that here are two key controlling factors: soil (defining the initial chemical form of P export) and hydrology (initiating P mo-bilisation); these are implicit in the source control vs. transport control argument outline above. These ideas are developed further by Haygarth and Jarvis (in press). A modified version of their conceptual framework is given in figure 17.1. It highlights the significance of hydrology as the driving mechanism of P transport. Title: 1.eps Creator: FreeHand 8.0 Preview:

This EPS picture was not saved with a preview included in it. Comment:

This EPS picture will print to a PostScript printer, but not to other types of printers.

Figure 17.1 Conceptual model of phosphorus transport

17.2 Background

(32)

ha*year. Approximately 57% of TP loss is derived from drained and undrained permanent grassland, with 23% from drained tilled land or grass leys and a further 18% from undrained tilled land or ley. Over the past 25 years, P inputs to land from fertilisers and manures have changed little but a greater proportion of tillage land is sown to winter cereals, more land is underdrained, and livestock density has increased (Withers, 1996). These land use trends have a potential to enhance P transport through soil erosion, subsurface P losses, and P enrichment of surface soils, respectively. The latter has re-ceived considerable attention because nutrient control is possible through regulation of fertiliser and manure inputs to land. Application of manures based on N demand results in overapplication of P because crop nutrient requirements are satisfied by a N:P ratio in the region 7-11:1 whilst manures generally fall in the range 2-6:1 (Smith et al., in press). Around 119 106 kg of P are returned annually to UK agricultural land as manures; an estimated 55% are applied to tillage land and 46% to grass-land (Burnhill et al., 1994; Smith et al., in press). Part of the explanation of the current UK P surplus of circa 10 kg per ha*year may lie in P enrichment of surface soils because livestock manure P is undervalued (Sharpley and Withers, 1994).

17.3 Flow pathways in agricultural catchments

Figure 17.2 illustrates the main hydrological pathways important in P transport at the hillslope scale. Key P inputs to the systems are indicated. This scale has been selected because it enables some inte-gration of current understanding of the mechanisms of P mobilisation with that of research on the magnitude of P loss. Title: 2.eps Creator: FreeHand 8.0 Preview:

(33)

17.3.1 Surface pathways

Two starting points for the generation of surface runoff are recognised. The first, infiltration-excess flow is generated when the infiltration capacity of the surface soil is exceeded, usually as a result of high intensity storm events. In the UK, rainfall intensities are generally low and the soil infiltration ca-pacity is unlikely to be exceeded (Kirkby, 1988) except where land management modifies the soil surface. Examples include intensive grazing of grassland or fodder crops (Heathwaite et al., 1989, 1990), which may generate infiltration-excess surface runoff on a field-wide scale. The second, satu-ration-excess surface runoff is topographically driven from spatially and temporally dynamic variable source areas (VSAs) (Beven and Wood, 1983). This pathway is triggered where the soil becomes saturated via lateral percolation above an impeding horizon. Saturation-excess surface runoff also occurs where the soil water table rises to the ground surface through convergent flow into hillslope hollows or where a rising stream water level result in saturation of near-stream zones. Under steady rainfall, saturation-excess flow requires much lower rainfall intensities to maintain it in comparison with infiltration-excess flow and is generally a more important mechanism of surface runoff generation. 17.3.2 Subsurface pathways

Subsurface flow may reach the drainage network via a number of pathways: (i) groundwater, (ii) lat-eral flow where soil layers have vertical conductivity < rainfall intensity, and (iii) where concave topographic contours create contributing areas because a high water table and/or subsurface imped-ance causes convergent flow. Where soils are deep and the bedrock permeable, percolation to groundwater rather than channelling of flow laterally will occur. The rate of subsurface flow depends on soil conductivity, which defines whether matrix flow (saturated/ piston flow) or preferential (mac-ropore) flow predominates. Preferential flow defines a rapid pathway of water transit through the soil. Certain antecedent thresholds (e.g. rainfall intensity and duration > 10 mm per day; soil moisture (θ) ≥ 0.3) must be satisfied before it occurs (Germann, 1986). It may occur naturally via soil macropores (Beven and Germann, 1982) or artificially via field drains (Armstrong and Garwood, 1991). Some soils, such as cracking clays, have a greater preponderance of macropores and hence more channel-ling of subsurface flow via this pathway.

17.4 Phosphorous fractionation in flows from agricultural land

Chemical fractionation procedures and soil P testing are considered by Edwards et al., (1997), Sims (1993, 1998), and Tunney et al. (1997). Recent research has focused on evaluation of P bioavailabil-ity in runoff (Dils and Heathwaite, 1998; Sharpley, 1993; Sharpley et al., 1992). Phosphorus is primarily mobilised as ions of inorganic orthophosphate or in association with organic or inorganic

(34)

ing the consequences of P transport for the quality of receiving waters. Their work suggests that or-ganic P fractions may form a larger part of transported P than previously thought and that attachment to soil colloids < 0.45µm may be particularly important.

Figure 17.3 indicates the main P fractions transported along the various hydrological pathways described in figure 17.2. In general, and primarily for tilled land, P transport in particulate form is as-sociated with surface runoff. Here the selective adsorption of P onto clay and silt-sized soil particles (as Fe/Al complexes or Ca/Mg phosphate) enables mobilisation with soil eroded from agricultural land. Transport of P in particulate organic form is important in grassland systems (Heathwaite et al., 1990). Subsurface pathways are commonly associated with P transport in dissolved form. However, preferential flow may also be an important pathway of particulate P transport (Dils and Heathwaite, 1996; Heathwaite, 1997) particularly attached to colloidal material (Haygarth et al., 1997). At the receiving end of the conceptual diagram (figure 17.3) factors such as mineral formation and dissolu-tion control P bioavailability (Lijklema, 1994).

Title: 3.eps Creator: FreeHand 8.0 Preview:

Figure 17.3 Phosphorus fractions transported in hillslope hydrological pathways

17.4.1 Surface pathways of P flow

Surface runoff has a strong affinity for P transport because the surface soil has the greatest effective depth of interaction (EDI) (Ahuja, 1986; Sharpley, 1985) and the highest concentrations of P (Hay-garth et al. 1998). Phosphorus residing in the surface 0.5 mm of soil appears to be most vulnerable

(35)

to export in runoff. Phosphorus transport in surface runoff is influenced by farming type, erosion po-tential, hydrologically effective rainfall, land use including fertiliser and manure amendments, the presence or absence of livestock, and soil total P (Chambers, 1997; Heathwaite, 1997). Surface runoff is important in physically transporting P via soil erosion (Sharpley and Smith, 1990). Even where erosion is minimal, elevated soil P can sustain high TP losses. For example, for grassland soils P transport in surface runoff may be exacerbated by high P concentrations at the soil surface as a result of organic matter inputs (Haygarth et al., 1996). Table 17.1 indicates the P fractionation in sur-face runoff from grassland in the Trent experimental catchment, Midlands, UK (Dils and Heathwaite, 1996). Here the concentration of dissolved P exceeds particulate P with most P transported in the DIP fraction. The large standard error indicates the wide spatial and temporal variation in TP trans-port in surface runoff. Similarly, Haygarth and Jarvis (1997) retrans-ported 70% TP transtrans-port in the dissolved fraction in surface runoff from grassland. Whilst Edwards and Daniel (1993) recorded TP loss in excess of 5 kg per hectare (95% dissolved P) from land receiving poultry litter. The rate, tim-ing and form of manure applications are important (Heathwaite et al., 1998) as is the time interval between application and rainfall (Haygarth and Jarvis, 1997; Hooda et al., 1996; Sharpley et al., 1994).

Table 17.1 Phosphorous fractionation (µg per l) in surface runoff from grassland (1994-1996) Trent catch-ment, Midlands, UK

Total P Total dissolved P Total particulate P

 

DIP DOP POP PIP

Mean 1,136 488 214 341 93

Standard error 77 61 38 32 17

N 60 60 60 14 14

TDP = DIP (dissolved inorganic P or molybdate reactive P) + DOP (dissolved organic P); TPP = POP (particulate organic P) + PIP (particulate inorganic P).

Source: Modified from Heathwaite (1997).

Where land management has increased the incidence of infiltration-excess surface runoff, sig-nificant transport of P may occur during storm events - often on a field-wide scale (Heathwaite, 1997). Some land management practices or crop types present a greater risk of P transport than oth-ers do. Where P transport is linked to soil erosion, high risk crops include winter cereals and winter vegetables, with temporary grass (< 5 years old), potatoes, sugar beet and maize of medium risk, and other arable crops such as spring cereals and oilseed crops of low risk (Chambers, 1997). In the

(36)

compact the soil surface and decrease the infiltration capacity of the surface soil leading to sheet ero-sion and associated P transport on a field-wide scale (Heathwaite et al., 1990). Chambers (1997) suggests that P loss by sheet erosion could be significant in the UK because, when triggered, it oper-ates over large land areas. Phosphorus transport in surface runoff from critical source areas (CSAs) has recently been shown important. Pionke et al. (1997), for example, suggest that 90% of the P load in receiving waters is derived from just 10% of the catchment. These CSAs are commonly linked to areas generating saturation-excess surface runoff. Partial source areas (PSAs) may also be effective at entraining P in surface runoff (Dils and Heathwaite, 1996). PSAs include effluent leakage from si-lage clamps, runoff from farmyards, channelling of flow along roads or tracks and tractor wheelings or animal tracks within fields. However, hydrological connectivity with the stream must exist for them to be significant factor in P transport to receiving waters. Although a number of researchers have rec-ognised the contribution from PSAs to P transport (Heathwaite et al., 1989; Dils and Heathwaite, 1996) it is difficult to quantify their actual contribution to the P load of receiving waters. Furthermore, the incidence of PSAs have a low frequency (although the P loss may be high) thus their impact re-mains under-researched.

17.4.2 Subsurface pathways of P flow

While the water reaching a stream via surface runoff largely constitutes rainfall falling during the event, subsurface flow reaching the channel is unlikely to be physically (or chemically) the same water as is actually falling as rainfall. Thus, evaluation of timelags in the system is crucial in understanding the mechanisms of P transport, especially as P transformations in transit through, for example, sorption of P from infiltrating water, will be far more important along subsurface pathways relative to surface pathways. To date there is little research on tracing P transformations during transit along subsurface pathways. This may be partly a reflection of the difficulty in isolating and measuring the P load along subsurface pathways and the perceived importance attached to surface pathways of P delivery during storm events. This assumption regarding the relative importance of surface vs. subsurface pathways of P transport may be well founded. Sharpley and Withers (1994), for example, compared P trans-port in surface runoff with losses in throughflow and artificial drainage and suggest that up to 9% of applied P fertiliser may be recorded in surface pathways compared to less than 1% in subsurface flow (although P loss in drainflow was higher). Hodgkinson and Withers (1996) demonstrated the importance of soil type, slope, and antecedent moisture on the incidence of surface vs. subsurface runoff and P transport. Their field losses of P in surface runoff and subsurface flow are presented in table 17.2a and 17.2b, respectively. Significant P transport (up to 1.76 kg TP per hectare) was re-corded in subsurface flow; although the contribution in this pathway remained smaller than that in surface runoff. Grassland clay soils record highest TP and DIP transport with the exception of losses from sandy soils during wet years.

Three subsurface pathways are recognised as having potential for P transport: first, near-surface lateral flow, owing to higher soil P concentrations in upper soil horizons - although P present in this horizon may not necessarily be mobile. For example, Chambers and Smith (1998) found that whilst soils receiving high loadings of organic manures and inorganic fertiliser showed P enrichment

(37)

in the upper 30 cm soil, there was no evidence of down profile mobilisation of P. Mobilisation ap-pears to be dependent on the mechanism of subsurface flow. In general, matrix flow is unlikely to initiate significant P transport, whereas preferential flow may be important (see below). The signifi-cance of P transport to groundwater by leaching depends on the depth to the water table and P loading at the soil surface. Smith et al. (1998), for example, report P enrichment of subsoil (> 45 cm depth) where freely draining soils have a history of high organic manure loadings. For P rich sites they found that the concentration of DIP in soil water moving in matrix flow below 30 cm increased where the Olsen extractable P concentration in the soil exceeded 70 mg per l. Thus, the potential for high groundwater loss of P exists where there is significant down profile transport of P in P rich sites where the groundwater table is shallow. Heckrath et al. (1995), for arable soils, report enhanced P loss in drainage water where the Olsen extractable P concentration in the plough layer exceeds 60 mg per kilogram. In the Netherlands, for example, the shallow water table and high P loading at the soil surface has created a high potential for P transport to groundwater. Here, van Riemsdijk et al. (1987) suggest that breakthrough of high P concentrations to groundwater are likely within 20-30 years if manure P loadings at the soil surface continue at current rates.

Table 17.2 Field losses of phosphorous (kg per ha) in surface runoff and subsurface flow for varying land use and soil types in England and Wales

(a) Surface runoff

Land use Soil type Slope Total Dissolved P (kg inorganic P per ha) (kg per ha)

Grassland clay 4° 3.30 1.37

Arable silt 5° 0.07 0.02

Arable sand 7° 0.17 (dry year) 0.01 (dry year) Arable sand 7° 9.33 (wet year) 0.29 (wet year)

(b) Subsurface flow

Land use Soil type Total Dissolved

P (kg inorganic P per ha) (kg per ha)

Arable clay 0.70 0.25

Arable clay 0.20 0.04

Grassland clay 1.76 0.39

(38)

Second, preferential flow may enable rapid subsurface transport of mobile P through soil macropores. Macropore flow reduces the time for interaction and hence the degree of transformation of P forms during transit. This may affect the bioavailability of P reaching the stream network. To date there has been little work on solute movement via preferential flow and no studies on movement of P with the exception of some initial work by Dils and Heathwaite (1996). The majority of studies that have inferred macropore flow have no direct evidence. Thus, this flow pathway is often assumed by a process of elimination. For example, Thomas et al. (1997) and Heckrath et al. (1995) suggested P transport via macropores in the silty clay loams of the Broadbalk plots at Rothamsted (Herts, UK) because high P concentrations were measured in tile-drain flow but the soils had a large adsorption potential and P was absent in soil solution at depth. In addition to dissolved P forms, this pathway may be important for P transport in particulate and colloidal form - particularly from grassland soils (Heathwaite, 1997). Dils and Heathwaite (1996) found around 68% TP transport in macropore flow from a mixed grass/arable catchment was in the particulate fraction, with mean concentrations of 842, 265 and 576 mg P l-1 for TP, TDP and TPP, respectively. Within the particulate fraction, the organic phase dominated, accounting for around 62% TP transported in macropore flow in the upper 45 cm soil.

Finally, artificial drainage acts like preferential flow to encourage rapid transit of water from land to stream. Approximately 6.4 million hectare of agricultural land have been underdrained in Eng-land and Wales: 71% on arable Eng-land and 28% on grassEng-land (Belding, 1971; Robinson and Armstrong, 1988). Phosphorus loss in drainflow is influenced by soil type (stability), soil total P, and excess winter rainfall (Chambers, 1997). Drained clay soils, for example, transmit water rapidly via cracks and mole channels; contact with subsoil is minimal and high P losses might be anticipated, par-ticularly where such soils receive high fertiliser or manure amendments. Dils and Heathwaite (in press) monitored the P fractionation in drainflow and streamflow for a number of storm events in the Trent catchment, Midlands, UK. The physico-chemical fractionation of TP appeared to be dependent on flow: at low flow DIP dominated and TP concentrations were low (<100µg per l), at high drainflow (>10 l per minute) associated with storm events, PP dominated with concentrations up to 1 mg TP per l. Kronvang et al. (1997) found that up to 18% of annual particulate P loss from a lowland arable catchment in Denmark was transported in subsurface drainage. Total P loss from grazed un-derdrained land with high animal manure inputs was over 5 times greater than unun-derdrained arable catchments (0.63 and 0.12 kg P per ha*year, respectively; Grant et al., 1996). A comparison of P export from drained and undrained agricultural land in England and Wales is given in table 17.3; P forms are not distinguished. Total P loss from agricultural land was estimated at 12,675 tonnes per year which is equivalent to 1.4 kg P per ha*year (Chambers, 1997). Underdrainage makes a signifi-cant contribution (38% of TP loss). The magnitude of loss via drainflow depends on the effective rainfall. Grassland makes the greatest contribution (43%) to TP loss although the P export coefficients are based on limited data. Lower P export around 0.5 kg per ha*year was recorded by Tunney et al. (1997) for Irish soils. Grassland drainage appeared to reduce the magnitude of P loss (Haygarth et al., 1998). Smith et al. (in press) recorded P transport in drainflow for tilled land receiving pig slurry, poultry litter or cattle FYM. The target rate of P application was 60 kg per hectare (range 37-103 kg P per hectare). Total P and DIP loss from pig slurry (1.15 kg TP per hectare; 0.44 kg DIP

(39)

per hectare) exceeded that of poultry litter or cattle FYM (ca. 0.25 kg TP per hectare; 0.05 kg DIP per hectare). The magnitude of P loss was correlated with peak drainflow with higher P concentra-tions recorded in the first drainage event following manure application. Thus concentrated liquid manures significantly increased P transport in drainflow with DIP concentrations in drainage waters up to 1,000 µg per l. In summary, drain flow represents an important potential pathway of P loss, especially in grassland catchments. For drained soils, where the likelihood of surface runoff has been reduced as a result of drainage, subsurface pathways of loss may represent the main pathway of P loss. It is also important to recognise that this pathway does not require the high magnitude, high in-tensity storm events necessary to generate surface runoff. Thus, it may generate higher P losses than previously recognised.

Table 17.3 Estimated phosphorous flow in surface runoff and drainflow

Land use Drainage Hydrologically Erosion Phosphorus Total status effective risk export P loss

rainfall (mm) coefficient (tonnes a-1) (kg h-1 a-1)

Permanent undrained < 200 - 0.7 568

grassland > 400 - 3.0 4,056

Permanent drained < 200 - 0.5 195

grassland > 400 - 2.0 1,296

Tillage undrained - very high 28.0 73

- high 6.0 204

- moderate 6.0 433

- slight 3.0 66

Tillage drained < 200 - 0.4 263

> 400 - 1.4 147

Source: Modified from Chambers (1997).

17.5 Conclusions and research needs

Table 17.4 presents a summary of the P 'signatures' recorded in different hydrological pathways for a mixed grassland/arable catchment in the Midlands, UK (after Dils and Heathwaite, in press). The data is a useful summary of the range and forms of P transport in different pathways. Highest P con-centrations were recorded in surface runoff and near-surface lateral flow in macropores (0-15 cm). However, the P signatures differed: surface runoff was dominated by the DIP fraction whilst P trans-port in shallow macropore flow was primarily in the particulate fraction. It is possible that rapid P

Agricultural data for life cycle assessments